KR19990044357A - Apparatus and method for transmitting and receiving digital information signal - Google Patents

Abstract

An apparatus for transmitting a digital information signal includes an input terminal for receiving the digital information signal, a first channel encoding unit for carrying out a first channel encoding step on an information word in a series of subsequent information words included in the digital information signal so as to obtain a channel word, a compression unit for carrying out a compression step on a channel word so as to obtain a compressed channel word, an error correction encoding unit for carrying out an error correction encoding of the compressed channel word so as to obtain a parity word, a second channel encoding unit for carrying out a second channel encoding step on the parity word so as to obtain a channel encoded parity word, and a formatting unit for combining the channel word and the channel encoded parity word into a composite transmission signal suitable for transmission via a transmission medium. With such apparatus, very long information words can be converted into channel words, while serious error propagation is avoided during transmission.

Description

Apparatus and method for transmitting and receiving digital information signal

The present invention relates to an apparatus for transmitting a digital information signal, an apparatus for receiving a digital information signal, and a method for transmitting a digital information signal. Such devices are known from EP 671,739 A2, document D1 of the related literature list. This apparatus has the form of an apparatus for recording a digital information signal on a recording medium such as a magnetic recording medium, and includes error correction means and channel encoding means.

When channel coding is simply defined, channel coding is a technique that realizes high transmission reliability while efficiently using channel capacity despite channel shortage. In essence, information theory suggests that a static channel can be made reliable by using a fixed portion of the channel for redundancy.

In transmission and recording / reproducing systems, source data is generally translated into two successive steps via (a) error correction code and (b) channel (or modulation) code, as described in Document D4 in the related literature list. Please see.

Error correction control is realized by adding extra symbols to the delivered message. These extra symbols allow the receiver to detect and / or correct some errors that may occur in the received message.

There are various classes of error correction codes. The most important class for recording related applications is the Reed-Solomon code (RS). The reason why this code is excellent for a recording / playback system, for example, is that it can solve not only burst errors but also random combinations. By using the margins present in the retrieved signal, the decoder can reconstruct the input sequence as accurately as possible.

The channel encoder performs a function of translating arbitrary user information and error correction symbols into a sequence following a given channel constraint. Examples of such constraints are spatial constraints or runlength constraints. The maximum information efficiency subject to channel input constraints is often referred to as the Shannon capacity of a noise-free channel subject to input constraints. Good code is code that achieves code efficiency close to the Shannon capacity of a constrained sequence, uses a simple implementation, and prevents the propagation of errors in the decryption process, and eventually, more practically, between these competing attributes. Corresponds to code with compromises.

Most existing write codes are byte oriented. In terms of channel capacity, the efficiency of such codes is typically less than 95%. According to the saying, "bigger is better", greater code efficiency can only be realized by using codes with very long code lengths, typically 500 to 1000 bits. Theoretically, we know how to construct such efficient code, but the biggest obstacle to the practical approach to this channel capacity is the large hardware problem for encoding and decoding, which means that the hardware size increases as the number of codewords used increases. That is, exponentially increasing with the length of the codeword.

By using algebraic coding techniques such as arithmetic, it is possible to realize a code whose complexity polynomically increases with codeword length. This algebraic encoding technique allows algorithms to be called to translate source words into code and vice versa, rather than performing translations into reference tables. See, for example, unpublished international patent application WO 96/00045, document D3 in the related literature list.

However, the second problem associated with using long code words is that the risk of error propagation is very high. One channel bit error can cause error propagation that destroys the entire decoded word. Of course, the longer the codeword, the greater the number of affected symbols.

After all, it is an object of the present invention to provide a digital information signal transmitter and a digital information signal receiver capable of preventing serious error propagation during a transmission process. The proposed signal processing for the transmitter and the proposed reverse signal processing for the receiver allows for the successful use of channel words with (very) long code words, thus achieving practical channel capacity. .

According to the invention, an apparatus for transmitting a digital information signal,

Input means for receiving a digital information signal;

First channel encoding means for performing a first channel encoding step on one of the consecutive information words included in the digital information signal to obtain a channel word;

Compression means for performing a compression step on the channel word to obtain a compressed channel word;

Error correction encoding means for performing a parity word on the compressed channel word;

Second channel encoding means for performing a second channel encoding step on the parity word to obtain a channel coded parity word;

And formatting means for synthesizing the channel word and the channel coded parity word into a composite transmission signal suitable for transmission through a transmission medium.

Also, the receiver is

Receiving means for receiving a composite transmission signal comprising a channel word and a corresponding channel coded parity word;

Demultiplexer means for retrieving a channel word and a corresponding channel coded parity word from said composite transmission signal;

Compression means for performing a compression step on said channel word to obtain a compressed channel word;

First channel decoding means for performing a channel decoding step on the channel coded parity word to obtain a parity word corresponding to the compressed channel word;

Error correction decoding means for performing an error correction step on the compressed channel word using the parity word to obtain a corrected compressed channel word;

Restoring means for performing a decompression step on the corrected and compressed channel words to obtain a decompressed channel word;

Second channel decoding means for performing a channel decoding step on the decompressed channel word to obtain an information word;

And output means for supplying the digital information signal in the form of a sequence of information words.

The present invention is based on the following idea.

First, Bliss proposed reversing the order of the conventional error correction coder and the channel coder, see reference D5 in the related literature list. In this Bliss encoding format, the constrained codeword (channel coded information word) is treated as binary input data of the error correction encoder in a general manner.

In a byte-oriented error control encoder (ECC) such as a conventional Reed-Solomon encoder, the constrained codeword (channel coded information word) is grouped into a plurality of bytes, and the generated parity word is constrained. To the end (or tip) of the codeword. The parity words generated in this way generally do not comply with certain constraints and are translated with the help of a second channel encoder. Accordingly, provisions must be made for connecting the various segments.

The corresponding decoding process in the above-described 'bleach method' is linear. First, the parity word is decoded using a corresponding channel code decoder. An error in the constrained sequence (a sequence of channel coded information words) may be corrected, and accordingly, the second channel decoder delivers a source sequence (sequence of information words). The efficiency of the channel code for the parity word is much smaller than the channel code for the information word. However, since the number of parity bits generally corresponds to a small portion of the number of input bits, the efficiency of the channel code for the parity word has a relatively small impact on the overall efficiency. It is very important that the error propagation of the channel code for converting the parity word is limited to a few bits, preferably one bit in a byte oriented system.

In the above-described 'bleach method', the constrained sequence becomes the input of the ECC encoder. Obviously, the constrained sequence is longer than the source data by a factor of 1 / R ₁ , where R ₁ represents the efficiency of the first channel encoder. In the following, it is assumed that the ECC can correct an error burst of a given length. Because ECC acts on a channel word, the corresponding number of user bytes it can correct is reduced by an R ₁ factor. For a recording system, this means that the burst error correction capability measured in geometric units, for example meters, is reduced by the same R ₁ factor. Second, the length of the constrained sequence instead of the user sequence must be less than the maximum imposed immediately by the RS code. The problem with the Bliss scheme described above is very serious, and despite its efficiency advantages, practical usefulness is limited for recording systems where correction of burst errors is an important requirement.

According to the present invention, these problems can be overcome by reconstructing the code and defining a third intermediate coding layer. Basically, the constrained sequence (i.e., the sequence of m-bit channel words) is compressed into a third intermediate sequence (sequence consisting of a plurality of compressed channel words) before being transmitted to the ECC encoder.

According to another aspect of the present invention, this compression process is performed by dividing the constrained sequence into a plurality of blocks consisting of p bits. This block length p is chosen such that the number of individual constrained sequences of length p is not greater than N, the field size of symbol error correction ECC. Accordingly, one-to-one mapping between p-tuple and ECC symbols can be defined. Using a small reference table, i.e. arithmetic, the p-tuple is translated in a unique way into an intermediate sequence of a plurality of bytes, the so-called compressed channel language. The number of bits of the compressed channel word generated as described above may be slightly larger than the number of bits of the original information word. On the other hand, it is used as the input of the ECC encoder of the intermediate sequence of compressed channel words, and a parity word is generated in a conventional manner. Note that the intermediate sequence is not translated. The parity word is modulated by a second constraint code (channel encoder) to obtain a channel coded parity word.

Sequences subordinate to this, i.e., constrained sequences involving constrained parity words, are transmitted or recorded.

The decoding process is performed in the following manner. First, a parity word is found by applying a channel decoding step to the channel coded parity word using a first channel decoder.

Next, the information word is found in the following way. Using the lookup table, the constrained sequence (sequence consisting of m-bit channel words) is translated into a sequence consisting of a plurality of bytes. This result can include an error, which can be corrected by the RS decoder.

Next, after the ECC decoding operation is finished, the corrected bytes are translated into a constrained sequence using the inverse of the lookup table. This corrected constraint sequence is then decoded by the second channel decoder.

Hereinafter, with reference to the preferred embodiments of the present invention together with the accompanying drawings will be described in more detail the above and further details of the present invention. In the drawings,

1 shows an embodiment of a transmission apparatus,

2A to 2F illustrate various words obtained at various positions inside the transmitting apparatus.

3 shows an embodiment of a receiving apparatus,

4 illustrates another embodiment of a formatting apparatus inside the apparatus shown in FIG. 1;

FIG. 5 shows another embodiment of a receiver inside the apparatus shown in FIG. 3.

1 shows an embodiment of a transmission apparatus according to the present invention. The apparatus has an input terminal 1 connected to an input 2 of the channel encoding circuit 4. The output 6 of the channel encoding circuit 4 is connected to the first input 8 of the formatting circuit 10 and the input 12 of the compression circuit 14. In addition, the output 16 of the compression circuit 14 is connected to the input 18 of the error correction coding circuit 20, and the error correction coding circuit is connected to the input 24 of the second channel coding circuit 26. Has an output 22. The output 28 of the channel encoding circuit 26 is also connected to the second input 30 of the formatting circuit 10.

Hereinafter, the operation of the embodiment shown in FIG. 1 will be described with reference to FIGS. 2A to 2F. The channel encoder 4 converts n-bit information words included in the information stream of the digital information signal given to the input terminal 1 to m-bit channel words. FIG. 2A schematically shows an n-bit information word given to the input 2 of the channel encoder 4, and FIG. 2B shows an m-bit supplied to the output 6 in response to the n-bit information word. Schematic representation of channel words. The length of the information word and the channel word is relatively longer than that normally performed in the channel encoder. As an example, n = 172 and m = 312. Thus, this channel encoder has an efficiency of about 0.55. The channel encoder circuit 4 converts a sequence of n-bit information words into a sequence of m-bit channel words that satisfy a particular (d, k) constraint. For example, d = 2 and k = 15.

The compression circuit 14 compresses each m-bit channel word into a corresponding compressed channel word consisting of a smaller number of bits. For example, the compressed channel word may have a length of 192 bits. 2c schematically illustrates such a compressed channel coded information word.

In this compression process, first, a channel-coded information word is divided into a plurality of subwords each having p bits. In the above embodiment in which m = 312, p can be selected as 13, so that 24 subwords exist in each of the channel coded information words. The compression circuit 14 then converts the subwords of each of these p bits into converted subwords of lower bit number r. For example, r = 8. This means that in the case of this embodiment, the compression circuit has a 13-to-8 bit converter. As a result, in the above embodiment, the compressed channel coded information word includes a sequence consisting of 24 8-bit transformed subwords.

As described above, the conversion to the transformed subword having a length of 8 bits has an advantage that error correction encoding can be performed on the compressed channel coded information word in a simple manner.

In this case, in order that the conversion from the p-bit subword to the r-bit converted subword is not ambiguous, a numerical value r is generated so that at most 2 ^r different ^subwords having a length p occur inside the channel coded information word. It should be noted that for p must select a number p which is the length of the subword within the channel coded information word. This p-bit subword to r-bit transformed subword can be performed using a lookup table.

The error correction encoding circuit 20 performs a known error correction encoding step for each compressed word to obtain a parity word as shown in FIG. 2D. This parity word may have a length of a number of bytes, for example 10 bytes. The channel encoding circuit 26 then converts this parity word into a channel coded parity word as shown in Fig. 2E. In this case, the channel encoding circuit 26 may have a form of an 8-to-15 bit converter. In addition, the sequence of channel coded parity words supplied by channel encoder 26 must meet the same (d, k) requirements as described above for channel coded information words.

The formatting circuit 10 synthesizes the channel coded information word and the corresponding channel coded parity word into one serial data stream as shown in Fig. 2F. In this case, it is obvious that the aforementioned (d, k) constraint must also be satisfied at the boundary between the channel coded information word and the channel coded parity word preceding or following it.

The synthesized serial data stream may be supplied to a transmission medium (TRM) as shown in FIG. 1 or recorded on a recording medium such as a magnetic recording medium.

3 shows an embodiment of a receiving apparatus according to the present invention. The apparatus includes a receiver 40 for receiving a transmission signal transmitted via a transmission medium (TRM). The output 42 of the receiver 40 is connected to the input 44 of the demultiplexer circuit 46. Further, the first output 48 of the demultiplexer 46 is connected to the input 50 of the compression unit 52, and the output 54 of the compression unit is connected to the first input 56 of the error correction decoding circuit 58. Connected. The demultiplexer 46 is provided with a second output 60 connected to the input 62 of the channel decoding circuit 64. In addition, the output 66 of the channel decoding circuit 64 is connected to the second input 68 of the error correction decoding circuit 58, and the error correction decoding circuit is connected to the input 72 of the recovery circuit 74. Output 70 is provided. In addition, the output 76 of the restoration circuit 74 is connected to the input 78 of the second channel decoding circuit 80. The output of the channel decoding circuit 80 is also connected to the output terminal 82.

Hereinafter, the operation of the embodiment shown in FIG. 3 will be described with reference to FIGS. 2A to 2F again. The data stream composed of the channel coded information word and the channel coded parity word indicated by (f ') is supplied to the receiver 40 and then to the demultiplexer 46. This data stream is referred to as (f ') as it refers to FIG. 2F, where the accent sign indicates that the data stream received and supplied to the demultiplexer 46 contains an error. Then, the received data stream is demultiplexer section 46, which consists of one data stream b 'consisting of channel coded information words fed to output 48 and another data consisting of channel coded parity words fed to output 60. Demultiplex into two separate data streams of stream e '. The data stream consisting of channel coded information words is then supplied to a compression circuit 52. In operation, this compression circuit 52 is exactly the same as the compression circuit 14 shown in FIG. Therefore, the compression circuit 52 compresses the m-bit channel coded information word into a compressed word in the same manner as described above with respect to the compression circuit 14. Thereafter, a data stream of channel coded parity words is supplied to the channel decoding circuit 64, which reconstructs the channel coded parity word into a data stream d 'consisting of the original parity word. As in the above embodiment, when the channel encoding circuit 26 is an 8-to-15 bit converter, the channel decoding circuit 64 becomes a 15-to-8 bit converter.

The error correction decoding circuit 58 then performs an error correction decoding process on the compressed word given to its input 56 in response to the corresponding parity word supplied to its input 68. Accordingly, the error correction and the compressed word are generated in the output 70 of the decoding circuit 58. This error corrected and compressed word is supplied to the reconstruction circuit 74. The channel decoding circuit 80 performs m-to-n bit conversion on the error-corrected m-bit compressed word to obtain an error-corrected n-bit information word. As in the above embodiment, when the channel encoding circuit 4 shown in FIG. 1 is a 172-to-312 bit converter, this channel decoding circuit 80 corresponds to a 312-to-172 bit converter.

As a result, a data stream composed of error-corrected information words is shown on the output terminal 82, which corresponds to a duplicate of the same data stream as is supplied to the input 1 of the transmitter.

FIG. 4 illustrates another embodiment of a formatter inside the apparatus shown in FIG. 1. The formatter 10 'shown in FIG. 4 has a multiplexer section 84 which multiplexes two data streams given at its inputs 8, 30 into a serial data stream, which serial data stream comprises a preamplifier ( 86. A serial data stream which is preamplified internally and then fed to a recording unit 88 having at least one recording head 90 and channel encoded on a recording medium 92, such as a magnetic recording medium or an optical recording medium. Will be recorded.

FIG. 5 shows another embodiment of the receiver 40 inside the apparatus shown in FIG. 3. The receiving section, indicated at 40 'in FIG. 5, has a reading section 96 for reading the serial data stream from the recording medium 92, which is amplified by the amplifier 98 and then output 42 Is supplied.

Although the present invention has been described with reference to preferred embodiments of the invention, it is obvious that these embodiments are not given to limit the invention. Accordingly, various modifications may be made by those skilled in the art to which the present invention pertains without departing from the scope of the invention as defined by the appended claims.

Moreover, the present invention encompasses all novel features and combinations of these features.

Further, reference numerals in the claims are given for the purpose of making the present invention easier to understand, and the scope of the present invention is not limited thereto.

Related literature

D1: EP 671,739 A2 (PHN 14.775)

D2: WO 95 / 24,775 A3 (PHN 14.774)

D3: international application no. WO 96/00045

D4: K. A. Schouhamer Immink, 'Coding techniques for digital recorders', Prentice-Hall Int. (UK) Ltd., New Jersey, 1991.

D5: W. G. Bliss, 'Circuitry for performing error correction calculations on baseband encoded data to eliminate error propagation', IBM Techn. Discl. Bill., Vol. 23, pp. 4633-4634, 1981.

Claims (13)

-Input means (1) for receiving a digital information signal, First channel encoding means (4) for performing a first channel encoding step on one information word of a series of consecutive information words included in said digital information signal to obtain a channel word, Compression means 14 for performing a compression step on the channel words to obtain a compressed channel word, Error correction encoding means (20) for obtaining a parity word by performing error correction encoding on the compressed channel word; Second channel encoding means (26) for performing a second channel encoding step on said parity word to obtain a channel encoded parity word; Formating means (10) for combining the channel word and the channel coded parity word into a composite transmission signal suitable for transmission over a transmission medium. The method of claim 1, The first channel encoding means 4 is configured to convert n-bit information words into m-bit channel words, where n and m are natural numbers and m is greater than n. Device. The method of claim 1, The second channel encoding means is configured to convert an r-bit parity word into an s-bit channel coded parity word, wherein r and s are natural numbers, and s is greater than r. Transmitter. The method of claim 1, And said formatting means further comprises recording means for recording said composite transmission signal in a track on a recording medium. Receiving means (40) for receiving a composite transmission signal comprising a channel word and a channel encoded parity word corresponding thereto; Demultiplexer means (46) for retrieving a channel word and a corresponding channel coded parity word from said composite transmission signal, Compression means (52) for performing a compression step on said channel word to obtain a compressed channel word, First channel decoding means (64) for performing a channel decoding step on the channel coded parity word to obtain a parity word corresponding to the compressed channel word; Error correction decoding means (58) for performing an error correction step on the compressed channel word using the parity word to obtain a corrected compressed channel word; Restoring means (74) for performing decompression on said corrected and compressed channel words to obtain decompressed channel words; Second channel decoding means (80) for performing a channel decoding step on the decompressed channel word to obtain an information word; An output means (82) for supplying said digital information signal in the form of a sequence of information words. The method of claim 5, The second channel decoding means 80 is configured to convert the m-bit decompressed channel word into an n-bit information word, where n and m are natural numbers and m is greater than n. Information signal receiver. The method of claim 5, The first channel decoding means is configured to convert an s-bit channel coded parity word into an r-bit parity word, where r and s are natural numbers, and s is greater than r. Receiver. The method of claim 5, And the receiving means further comprises reading means for reading a composite transmission signal from a track on a recording medium. The method according to claim 1 or 5, and n and m are greater than 100. The method according to claim 1 or 5, The compression means, dividing means for dividing the m-bit channel word into m / p subwords each composed of p bits; Converting means for converting the p-bit subwords into r-bit transformed subwords such that the m / p transformed subwords form the compressed channel word, wherein p and wherein r is an integer and p is greater than r. The method of claim 10, p is selected such that the maximum of different p-bit subwords that can occur in a sequence of a plurality of m-bit channel words is equal to at most 2 ^r . Receiving a digital information signal; Performing a first channel encoding step on one information word of a series of consecutive information words included in the digital information signal to obtain a channel word; Compressing the channel word to obtain a compressed channel word, Error correcting encoding the compressed channel word to obtain a parity word; Performing a second channel encoding step on the parity word to obtain a channel coded parity word; And synthesizing the channel word and the channel coded parity word into a composite transmission signal suitable for transmission over a transmission medium. The method of claim 12, And said synthesizing step further comprises a substep of recording said synthesized transmission signal on a track on a recording medium.